Proposing a New Speech Enhancement Method based on Spectral Subtraction and Binary Masking With a Bank of 128 Gamma-tone Filters

In order to the enhancement of the quality of speech corrupted by additive noise, a speech enhancement method has been put forward based on the combination of spectral subtraction and binary masking. Spectral subtraction is a powerful method for removing noise from speech and binary masking provides essential elements to be used in monaural speech segregation. In the proposed combined method, first, spectral subtraction is used for reduction of noise in noisy speech and then binary masking is used for monaural speech segregation from musical noise introduced by spectral subtraction. The binary masking method, isolates the basic principle of target voice from other signals by using time frequency decomposition, energy masking and unites grouping. This masking is like what the human ear does in noisy environment. In our implementation for binary masking, an auditory filter (Gamma-tone) is divided into different frequency sub-bands. From these sub-band channels, channels 1,2,4,8,16,32,64,128 have been used from this bank of 128 Gamma-tone filter for implementing the binary masking. Evaluations show that the proposed combined method can improve the signal to noise ratio from 5 to 19 db for experimented signals and have better performance than binary masking or spectral subtraction in most situations, especially when noise and speech have not similar power spectrum.

Keywords:

speech enhancement , spectral subtraction , binary masking , Gamma-tone filter bank , musical noise